Semiconductor devices are widely used in many products. An important trend of developing semiconductor device is the miniaturization of the products using high density semiconductor devices. However, competing with this need for smaller dimensions is the increasing need for functionality that the consumers of semiconductor devices demand. The increase in functionality tends to increase the size and complexity of the semiconductor devices and the number of semiconductor devices per module. This trend consequently results in area distributed I/O terminals and corresponding C.sup.4 area bonding technique.
Another significant factor is maintaining low cost of manufacturing despite the increase in semiconductor device complexity and density. One significant cost in manufacturing a semiconductor device is the lead frame. As a result, direct chip attach (DCA) type assemblies are gaining popularity. In DCA, the pads of a semiconductor device are directly attached to a substrate, such as PCB. Thus the cost and size of an individual package for the semiconductor device is reduced.
While C.sup.4 (controlled collapse chip connection) and DCA techniques becomes popular, the problem with thermal coefficient of expansion (TCE) mismatch between chip and substrate carrier becomes serious in particular with the larger integrated circuit (IC) chips, high TCE, low-cost organic substrate, and smaller solder joints. Due to the TCE mismatch between the silicon IC (2.5 ppm) and low cost organic substrate, in particularly, the FR-4 (18-24 ppm) printed wiring board, the temperature cycling excursions generate thermomechanical stresses to the solder joints and result in performance degradation of packaged systems.
Underfill encapsulant is used to reinforce physical, mechanical and electrical properties of the solder joints connecting the chip and the substrate. The encapsulant does not only provide dramatic fatigue life enhancement, but also extends its use to a variety of organic and inorganic substrate materials, resulting in ten to hundred fold improvement in fatigue life as compared to an unencapsulated package. Therefore, the new technique of underfill encapsulation has been gaining increasing acceptance.
However, more than 90% of current underfill encapsulants dispense liquid encapsulants on one or two edges of an assembled flip-chip package. This allows capillary action to draw the underfill into the gap between the chip and substrate of the assembled package to complete the encapsulation process. This process has two disadvantages: (1) The process of reflowing solder bump and the process of underfilling and curing the encapsulants are separated, which result in lower production efficiency. (2) The limit of capillary force results in the limit of flow distance for underfill materials, which further limits the chip size. As such it becomes a production bottleneck.
A no-flow underfilling process was invented to dispense the underfill materials on the substrate or the semiconductor devices at first, then perform the solder bump reflowing and underfill encapsulant curing simultaneously. Therefore, no-flow underfilling process not only eliminates the strict limits on the viscosity of underfill materials and package size, but also improves the production efficiency. Pennisi et al, in one of their patents [U.S. Pat. No. 5,128,746], described the similar process as the no-flow underfilling process and herein cited as a reference. Till now, however, no-flow underfilling process has not been practically used in industrial mass production. The reason mainly lies in lacking of a successful no-flow underfill material.
A successful no-flow underfill encapsulant should meet primary requirements: (1) minimal curing reaction should occur at the temperature below solder bump reflow temperature (.about.190-230.degree. C.). (2) rapid curing reaction should take place after the maximum solder bump reflow temperature. (3) good adhesion to passivation layer, chip, substrate, solder mask, and solder joints. (4) reasonable shrinkage of the material during curing. (5) low TCE. (6) self-fluxing ability. (7) exhibit a reasonable modulus to minimize the residual thermal stress resulted from the curing process and sequence temperature cycling condition.
Silica filled epoxy resins were most widely investigated candidates due to their low cost and good integrated properties. There are many kinds of epoxy resins in current market, and the formulations based on these epoxy resins and protected by patents are enormous. Most of these formulations are based on aromatic epoxy resin such as bisphenol A type epoxy resin and bisphenol F type epoxy resin, and the Tg of these formulations after curing generally ranges from 100 to 140.degree. C. Polynuclear aromatic epoxides are now under developing, especially in Japan, and the Tg of cured polynuclear aromatic epoxides generally ranges from 180 to 220.degree. C., which is close to the Tg of polyimide (250.degree. C.). Some research results of polynuclear aromatic epoxides were reported in "Polymeric Materials for Microelectronics Applications" edited by Hiroshi Ito, Seiichi Tagawa, and Kazuyuki Horie. However, most of polynuclear aromatic phenols are cancer suspending chemicals and the cost of producing them are relatively high. As such, they are not widely used in the industries. Cycloaliphatic epoxides has been commercialized many years ago and its market price is relatively inexpensive. Cycloaliphatic epoxide attracts increasing attentions due to its lacking of phenylene UV chromophore, low concentration of mobile chlorine ion, and low viscosity. Papathomas et al in one of their patents [U.S. Pat. No. 5,194,930] developed cycloaliphatic epoxide/anhydride system for solder interconnection structure application in flip-chip technology and herein cited as a reference. But they did not mention any catalyst, fluxing agents and viscosity controlling agents suitable for no-flow underfill formulations. Instead, they used glycol as curing accelerator with curing peak temperature around 160.degree. C. which is out of useful range for no-flow underfill applications. Pennisi et al, in one of their patents [U.S. Pat. No. 5,128,746] developed an epoxy based material for their processing, but they did not realize the importance of the viscosity controlling agents and silane coupling agents for no-flow underfilling process. Moreover, their material contains organic acid type fluxing agent, which is corrosive to the solder interconnects.